JOURNAL OF FINANCE & CORPORATE GOVERNANCE
Volume 1, Numéro 2, Pages 07-23
2017-12-30

A New Hybrid Expansion Function Based Mutual Information For A Multilayer Neural Networks Optimization



Authors : Kais Ncibi . Djenina Amor . Sadraoui Tarek . Faycel Mili .

Abstract

Function expansion was used to expand initial features based on a non linear transformation. Many known expansion functions are found such the trigonometric, the polynomial, the Legendre polynomial, the power series, the exponential and the logarithmic transformation. This paper present a comparison between different expansion functions based on mutual information and different performance functions. We propose a new expansion process able to improve the correspondent mutual information and the final performance. The process was tested; using different benchmark databases, and shows his ability to improve results of classification problems

Keywords

Function expansion; multilayer perceptron, Classification, mutual information, features selection.

Les articles similaires

Intrusion Detection Based On Neural Networks Neural Networks In Iot

Chaima Bensaid . 
pages 29-33.

Towards The Proposition Of A New Approach Based On Artificial Neural Networks : Application To Quality Of Life Prediction Through Illness Perception With Chronic Illness Patients

Aberkane Salah . 
pages 98-103.

Conception Of Robust Neural Networks To Improve Hybrid Control Of An Induction Motor

Laribi M. .  Barazane L. .  Larbès C. .  Malek A. . 
pages 131-144.

Automatic Diagnosis System For Classification, Segmentation, Detection And Tracking Of Skin Lesion Based On Deep Convolutional Neural Networks

Benyahia Samia .  Meftah Boudjelal . 
pages 23-27.

Direct Power Control Based Artificial Neural Networks Of Doubly Fed Induction Generator For Wind Energy Conversion System Application

Djeriri Youcef . 
pages 592-603.

Formats de citation